minio/warp

How minio/warp works

Warp is an S3 benchmarking tool with CLI-driven object workloads, distributed client execution, live/final result analysis, and Iceberg catalog/table benchmarks. The usual flow configures a benchmark, executes it against S3 (locally or through clients), then aggregates and presents results. The graph includes documented benchmark branches and sampled internal layers; unsampled command implementations are represented only at their exact paths, and their specific call paths are not inferred.

Main entry points: aggregate.go Result aggregation · cli.go Command interface · cli S3 workloads · benchserver.go Run coordinator

Components 26

CLI workflows

Program entry main.go
Command interface cli/cli.go
Flags and config cli/flags.go

Benchmark execution

S3 workloads cli
S3 client setup cli/client.go
HTTP transports cli/client_transport.go
Data generation pkg/generator/generator.go

Distributed runs

Client protocol cli/benchclient.go
Run coordinator cli/benchserver.go

Results and analysis

Result aggregation pkg/aggregate/aggregate.go
Live updates pkg/aggregate/live.go
Request statistics pkg/aggregate/requests.go
Throughput metrics pkg/aggregate/throughput.go
Analysis commands cli/analyze.go

Iceberg benchmarks

Iceberg workloads cli
Catalog operations pkg/iceberg/catalog.go
Table commits pkg/iceberg/commit.go
Dataset generation pkg/iceberg/dataset.go
Parquet encoding pkg/iceberg/parquet.go

Result presentation

Monitor API api/api.go
Live web UI wui/server.go
UI data handler wui/handlers.go

Other components

Operator
S3 service
Warp clients
Browser

Connections 30

Operator → Program entry: runs command README.md
Program entry → Command interface: starts CLI main.go
Operator → Command interface: selects benchmark README.md
Command interface → Flags and config: uses options cli/flags.go
Command interface → S3 workloads: dispatches workload
S3 workloads → S3 client setup: uses S3 client cli/client.go
S3 client setup → HTTP transports: builds transport cli/client.go
S3 client setup → S3 service: connects to README.md
S3 workloads → Data generation: uses generated data
Command interface → Iceberg workloads: selects Iceberg run README.md
Iceberg workloads → Catalog operations: runs catalog operations cli/iceberg_catalog_read.go
Iceberg workloads → Table commits: generates commits README.md
Iceberg workloads → Dataset generation: writes table data cli/iceberg_sustained.go
Dataset generation → Parquet encoding: uses encoding
Operator → Run coordinator: starts distributed run README.md
Run coordinator → Warp clients: coordinates clients README.md
Warp clients → Client protocol: accepts run requests cli/benchclient.go
Client protocol → S3 workloads: executes benchmark cli/benchclient.go
Run coordinator → Result aggregation: merges results cli/benchserver.go
S3 workloads → Result aggregation: records operations
Result aggregation → Live updates: updates live data pkg/aggregate/aggregate.go
Result aggregation → Request statistics: computes request stats pkg/aggregate/aggregate.go
Result aggregation → Throughput metrics: computes throughput pkg/aggregate/aggregate.go
Operator → Analysis commands: requests analysis README.md
Analysis commands → Result aggregation: reads results cli/analyze.go
Run coordinator → Monitor API: serves run status cli/benchserver.go
Monitor API → Live updates: requests updates api/api.go
Live web UI → UI data handler: registers routes wui/server.go
UI data handler → Live updates: polls live data wui/handlers.go
Browser → Live web UI: views results

Drawn by GitDiagram on November 11, 2026 from the repository's file tree, README and source files. Read as Markdown ·